Hi All,
\nI've been having trouble with getting a score back, the notebook runs fine, is offline, etc, and I have a submission.csv with the 3 patient_ids as well as the 14 columns for the injuries. Would appreciate if anyone could help. Linking public notebook - https://www.kaggle.com/code/jdonnelly0804/rsna-first-submission
", "messages": [ { "id": 2413459, "postDate": "2023-08-29T00:02:01.970Z", "content": "Hi All,
\nI've been having trouble with getting a score back, the notebook runs fine, is offline, etc, and I have a submission.csv with the 3 patient_ids as well as the 14 columns for the injuries. Would appreciate if anyone could help. Linking public notebook - https://www.kaggle.com/code/jdonnelly0804/rsna-first-submission
", "rawMarkdown": "Hi All,\n\nI've been having trouble with getting a score back, the notebook runs fine, is offline, etc, and I have a submission.csv with the 3 patient_ids as well as the 14 columns for the injuries. Would appreciate if anyone could help. Linking public notebook - https://www.kaggle.com/code/jdonnelly0804/rsna-first-submission", "votes": 3 }, { "id": 2418629, "postDate": "2023-09-01T12:36:50.690Z", "content": "In my case it was due to corrupt dicom file mentioned here.
\nI was reading files in nested loops and when that file was encountered, loop was throwing an error so a partial submission file was created later due to missing scans.
", "rawMarkdown": "In my case it was due to corrupt dicom file mentioned [here](https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/435815).\n\nI was reading files in nested loops and when that file was encountered, loop was throwing an error so a partial submission file was created later due to missing scans.", "votes": 1 }, { "id": 2415331, "postDate": "2023-08-30T08:47:17.350Z", "content": "As a result of experiments, I have established the following.
\nI was writing about 10 MB of files to the '/kaggle/working' directory. Everything goes fine.
\nAnd if you write a lot (About 100 MB), I did not set the exact number, then at the end there is a CSV Not Found error.
\nI don't presume to judge why, but it's a fact. I didn't find any information on this issue in Google.
Interesting… Thanks for the heads up!
", "rawMarkdown": "Interesting... Thanks for the heads up!" }, { "id": 2418025, "postDate": "2023-09-01T04:44:53.487Z", "content": "I got the same error but I don't think this would be the reason.
", "rawMarkdown": "I got the same error but I don't think this would be the reason." } ] }, { "id": 2414197, "postDate": "2023-08-29T13:13:03.927Z", "content": "I have problems with this error too \"Submission CSV not found\"
\nThe simplest output of the file passes without an error. As soon as I added a little logic to the code, it issues a Submission CSV not found. Despite the fact that it is under the try catch block and in theory should be executed in any case.
\nAlthough, my code writes intermediate files to the 'working' directory for convenience. Although yesterday I conducted experiments with writing files to the working directory and even reading them. Everything worked fine. What doesn't work in my new code - now I can only guess. Attempts are over for today. I will report back when I find the cause.
\nBy the way, have you tried switching gpu P100 to t4 or vice versa?
I haven't! I'll give it a shot today, and I'll also let you know if I figure out this submission error… sure is a pain, no idea what's wrong! My output looks exactly the same as other notebooks I've looked at.
", "rawMarkdown": "I haven't! I'll give it a shot today, and I'll also let you know if I figure out this submission error... sure is a pain, no idea what's wrong! My output looks exactly the same as other notebooks I've looked at.", "replies": [ { "id": 2414345, "postDate": "2023-08-29T15:11:51.517Z", "content": "Do you have an example of minimal code that outputs \"CSV not found\"?
\nYour link to the notebook returns a 404 error, I would like to see if there are any similarities with my code.
\"My output looks exactly the same as other notebooks I've looked at.\"
\nMy notebook also outputs the CSV file during normal execution, but it cannot find it when submitted. This suggests that the error is not related to the output, but rather to an issue in the code that works fine during normal notebook execution but stops the code from running during submission.
My notebook produces warnings, which may be related to the issue.
\n [WARN:0@1949.697] global loadsave.cpp:248 findDecoder imread_('/kaggle/working/presave/9537-1117/0.png'): can't open/read file: check file path/integrity.
If I don't find better ideas, I will investigate this version.
", "rawMarkdown": "Do you have an example of minimal code that outputs \"CSV not found\"? \nYour link to the notebook returns a 404 error, I would like to see if there are any similarities with my code. \n\n\"My output looks exactly the same as other notebooks I've looked at.\"\nMy notebook also outputs the CSV file during normal execution, but it cannot find it when submitted. This suggests that the error is not related to the output, but rather to an issue in the code that works fine during normal notebook execution but stops the code from running during submission. \n\nMy notebook produces warnings, which may be related to the issue.\n [WARN:0@1949.697] global loadsave.cpp:248 findDecoder imread_('/kaggle/working/presave/9537-1117/0.png'): can't open/read file: check file path/integrity. \n \nIf I don't find better ideas, I will investigate this version." }, { "id": 2414634, "postDate": "2023-08-29T19:19:15.563Z", "content": "preds, _ = learn.get_preds(dl=test_dl)\nsubmission = pd.read_csv(\"/kaggle/input/rsna-2023-abdominal-trauma-detection/sample_submission.csv\")\nsubmission[df.columns]= preds\nsubmission.to_csv('submission.csv', index=False)
I don't get that warning however. For now, I just created a separate notebook that reads in the submission DF, and then merges in my predictions from my training notebook… for whatever reason that works when I submit, and at least I can get a score and iterate for now. Also, I tried using the T4, and found that I was able to train slightly faster.
", "rawMarkdown": "`preds, _ = learn.get_preds(dl=test_dl)\nsubmission = pd.read_csv(\"/kaggle/input/rsna-2023-abdominal-trauma-detection/sample_submission.csv\")\nsubmission[df.columns]= preds\nsubmission.to_csv('submission.csv', index=False)`\n\nI don't get that warning however. For now, I just created a separate notebook that reads in the submission DF, and then merges in my predictions from my training notebook... for whatever reason that works when I submit, and at least I can get a score and iterate for now. Also, I tried using the T4, and found that I was able to train slightly faster." } ] } ] } ], "comments": [ { "id": 2418629, "author_name": "Gunes Evitan", "author_url": "", "post_date": "2023-09-01T12:36:50.690000", "content": "In my case it was due to corrupt dicom file mentioned here.
\nI was reading files in nested loops and when that file was encountered, loop was throwing an error so a partial submission file was created later due to missing scans.
", "votes": 1, "replies": [] }, { "id": 2415331, "author_name": "Anuar-s", "author_url": "", "post_date": "2023-08-30T08:47:17.350000", "content": "As a result of experiments, I have established the following.
\nI was writing about 10 MB of files to the '/kaggle/working' directory. Everything goes fine.
\nAnd if you write a lot (About 100 MB), I did not set the exact number, then at the end there is a CSV Not Found error.
\nI don't presume to judge why, but it's a fact. I didn't find any information on this issue in Google.
Interesting… Thanks for the heads up!
", "votes": 0, "replies": [] }, { "id": 2418025, "author_name": "Gunes Evitan", "author_url": "", "post_date": "2023-09-01T04:44:53.487000", "content": "I got the same error but I don't think this would be the reason.
", "votes": 0, "replies": [] } ] }, { "id": 2414197, "author_name": "Anuar-s", "author_url": "", "post_date": "2023-08-29T13:13:03.927000", "content": "I have problems with this error too \"Submission CSV not found\"
\nThe simplest output of the file passes without an error. As soon as I added a little logic to the code, it issues a Submission CSV not found. Despite the fact that it is under the try catch block and in theory should be executed in any case.
\nAlthough, my code writes intermediate files to the 'working' directory for convenience. Although yesterday I conducted experiments with writing files to the working directory and even reading them. Everything worked fine. What doesn't work in my new code - now I can only guess. Attempts are over for today. I will report back when I find the cause.
\nBy the way, have you tried switching gpu P100 to t4 or vice versa?
I haven't! I'll give it a shot today, and I'll also let you know if I figure out this submission error… sure is a pain, no idea what's wrong! My output looks exactly the same as other notebooks I've looked at.
", "votes": 0, "replies": [ { "id": 2414345, "author_name": "Anuar-s", "author_url": "", "post_date": "2023-08-29T15:11:51.517000", "content": "Do you have an example of minimal code that outputs \"CSV not found\"?
\nYour link to the notebook returns a 404 error, I would like to see if there are any similarities with my code.
\"My output looks exactly the same as other notebooks I've looked at.\"
\nMy notebook also outputs the CSV file during normal execution, but it cannot find it when submitted. This suggests that the error is not related to the output, but rather to an issue in the code that works fine during normal notebook execution but stops the code from running during submission.
My notebook produces warnings, which may be related to the issue.
\n [WARN:0@1949.697] global loadsave.cpp:248 findDecoder imread_('/kaggle/working/presave/9537-1117/0.png'): can't open/read file: check file path/integrity.
If I don't find better ideas, I will investigate this version.
", "votes": 0, "replies": [] }, { "id": 2414634, "author_name": "jdonnelly0804", "author_url": "", "post_date": "2023-08-29T19:19:15.563000", "content": "preds, _ = learn.get_preds(dl=test_dl)\nsubmission = pd.read_csv(\"/kaggle/input/rsna-2023-abdominal-trauma-detection/sample_submission.csv\")\nsubmission[df.columns]= preds\nsubmission.to_csv('submission.csv', index=False)
I don't get that warning however. For now, I just created a separate notebook that reads in the submission DF, and then merges in my predictions from my training notebook… for whatever reason that works when I submit, and at least I can get a score and iterate for now. Also, I tried using the T4, and found that I was able to train slightly faster.
", "votes": 0, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "2413459": "Hi All,\n\nI've been having trouble with getting a score back, the notebook runs fine, is offline, etc, and I have a submission.csv with the 3 patient_ids as well as the 14 columns for the injuries. Would appreciate if anyone could help. Linking public notebook - https://www.kaggle.com/code/jdonnelly0804/rsna-first-submission", "2418629": "In my case it was due to corrupt dicom file mentioned [here](https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/435815).\n\nI was reading files in nested loops and when that file was encountered, loop was throwing an error so a partial submission file was created later due to missing scans.", "2415331": "As a result of experiments, I have established the following.\nI was writing about 10 MB of files to the '/kaggle/working' directory. Everything goes fine.\nAnd if you write a lot (About 100 MB), I did not set the exact number, then at the end there is a CSV Not Found error.\nI don't presume to judge why, but it's a fact. I didn't find any information on this issue in Google.", "2414197": "I have problems with this error too \"Submission CSV not found\"\n\nThe simplest output of the file passes without an error. As soon as I added a little logic to the code, it issues a Submission CSV not found. Despite the fact that it is under the try catch block and in theory should be executed in any case.\n\nAlthough, my code writes intermediate files to the 'working' directory for convenience. Although yesterday I conducted experiments with writing files to the working directory and even reading them. Everything worked fine. What doesn't work in my new code - now I can only guess. Attempts are over for today. I will report back when I find the cause.\nBy the way, have you tried switching gpu P100 to t4 or vice versa?\n" } }